Construction of Giant Dendrimers Using a Tripodal Building Block

Construction of Giant Dendrimers Using a Tripodal Building Block ..... Gold nanoparticle-cored poly(propyleneimine) dendrimers as a new platform for ...
0 downloads 0 Views 252KB Size
Published on Web 05/21/2003

Construction of Giant Dendrimers Using a Tripodal Building Block Jaime Ruiz,† Gustavo Lafuente,† Sylvia Marcen,† Catia Ornelas,† Sylvain Lazare,‡ Eric Cloutet,§ Jean-Claude Blais,⊥ and Didier Astruc*,† Contribution from the Laboratoire de Chimie Organique et Organome´ tallique, UMR CNRS No. 5802, Laboratoire de Physico-Chimie Mole´ culaire, UMR CNRS 5803, Laboratoire de Chimie des Polyme` res Organiques, UMR CNRS 5629, UniVersite´ Bordeaux I, 33405 Talence Cedex, France, and Laboratoire de Chimie Structurale Organique et Biologique, UMR CNRS No. 7613, UniVersite´ Paris VI, 4 Place Jussieu, 75252 Paris, France Received September 4, 2002; E-mail: [email protected]

Abstract: Giant pentane-soluble organo-silicon dendrimers have been synthesized using a triallylphenol brick according to a new divergent construction that uses a hydrosilylation-nucleophilic substitution sequence up to the ninth generation (G9). All the reactions were monitored by 1H, 13C, and 29Si NMR until G9, indicating that they were clean at the NMR accuracy until this last generation. MALDI TOF mass spectra were recorded for G1 to G4 and show the nature and amounts of defects that are intrinsic to the divergent construction. This technique and SEC (recorded up to G5) confirm the monodispersity (1.00 to 1.02) from G1 to G5. HRTEM and AFM images recorded for the high generations disclose the expected smooth dendrimer size progression and the globular shape. At G9, the theoretical number of termini (TNT) is 177 407 branches (abbreviation: G9-177 047). It is estimated that more than 105 terminal branches are actually present in the G9 dendrimer, far beyond the De Gennes “dense-packing” limit (6000 branches), and it is believed that the branch termini turn inside the dendrimer toward the core. This is corroborated by lower reaction rates and yields for the highest generation numbers presumably due to intradendritic reactions. It is probable that the dendritic construction is limited by the density of branches inside the dendrimer, i.e., far beyond the dense-packing limit.

Introduction

Dendrimers have numerous potential applications in biology, catalysis, and materials science.1-8 This field of molecular science is now well spread, yet only very few syntheses and characterizations of very large dendrimers have been reported. The largest dendrimers whose characterizations have been †

LCOO, Universite´ Bordeaux I. LPCM, Universite´ Bordeaux I. LPCO, Universite´ Bordeaux I. ⊥ LCSOB, Universite ´ Paris VI. ‡ §

(1) (a) Tomalia, D. A.; Taylor, A. N., Goddard, W. A., III. Angew. Chem. Int. Ed. Engl. 1990, 29, 138-175. (b) Fre´chet, J. M. J. Science, 1994, 263, 1710-1715. (c) Ardoin, N.; Astruc, D. Bull. Soc. Chim. Fr. 1995, 132, 875-909. (2) (a) Newkome, G. R.; Moorefield, C. N.; Vo¨gtle, F. Dendrimers and Dendrons: Concepts, Syntheses and Applications; VCH: Weinheim, 2001. (b) AdVances in Dendritic Molecules; Newkome, G., Ed.; JAI: Greenwich, CT, Vols. 1-5, 1994-1998 and 2002. (c) Dendrimers, Topics in Current Chemistry; Vo¨gtle, F., Ed.; Springer-Verlag: Berlin, Vol. 1, 1997; and Vol 2, 2000. (d) Dendrimers and Other Dendritic Polymers; Tomalia, D., Fre´chet, J. M. J., Eds.; Wiley-VCH: New York, 2002. (3) Zeng, F.; Zimmerman, S. C. Chem. ReV. 1997, 97, 1681-1712. (4) Matthews, O. A.; Shipway, A. N.; Stoddart, J. F. Prog. Polym. Sci. 1998, 23, 1-56. (5) Venturi, M.; Serroni, S.; Juris, A.; Campagna, S.; Balzani, V. In Dendrimers, Topics in Current Chemistry; Vo¨gtle, F., Ed.; Springer-Verlag: Berlin, 1998; Vol. 197, pp 193-228. (6) (a) Bosman, A. W.; Janssen, H. M.; Meijer, E. W. Chem. ReV. 1999, 99, 1665-1688. (b) Sijbesma, R. P.; Meijer, E. W. Curr. Opin. Coll. Int. Sci. 1999, 4, 24-34. (c) Weener, J.-W.; Baars, M. W. P. L.; Meijer, E. W. In ref 2d, pp 387-424. (d) Bosman, A. W.; Bruining, M. J.; Kooijman; Spek, A. L.; Janssen, R. A. J.; Meijer, E. W. J. Am. Chem. Soc. 1998, 120, 55478548. 7250

9

J. AM. CHEM. SOC. 2003, 125, 7250-7257

published9-11 have total numbers of termini (TNT) up to about 3000. One report appeared for large dendrimers with up to more than 2 × 106 branches, but these dendrimers were not characterized by spectroscopy, mass spectra, or microscopy beyond the first generation, and their construction might have been marred by cross-linking.12 Other large dendrimer constructions were not larger than the dense-packing limit.13 Two groups have reported that dendrimers terminated by amino groups become insoluble when the TNT reaches the dense-packing limit (TNT: 30729 and 614411). These observations seemingly confirm De Gennes’ dense-packing theory13 set for PAMAM dendrimers according to which the steric congestion occurring at the dendritic surface inhibits further regular dendrimer (7) (a) Hecht, S.; Fre´chet, J. M. J. Angew. Chem., Int. Ed. 2001, 40, 74-91. (b) Fre´chet, J. M. J. Chem. ReV. 2001, 101, 3819-3867. (c) Aliphatic ester dendrimers up to 30 711 Da and 192 masked hydroxyl groups (in six generations) have recently been conveniently prepared in high yield and purity: Ihre, H.; Padilla De Jesu´s, O. L.; Fre´chet, J. M. J. J. Am. Chem. Soc. 2001, 123, 5908-5917. (8) Crooks, R. M.; Zhao, M.; Sun, L.; Chechik, V.; Yeung, L. K. Acc. Chem. Res. 2001, 34, 181-190. (9) Tomalia, D. A.; Baker, H.; Dewald, J. R.; Hall, M.; Kallos, G.; Martin, S.; Roeck, J.; Ryder, J.; Smith, P. Macromolecules 1986, 19, 2466-2470. (10) van der Made, A. W.; van Leeuwen, P. W. N. M.; de Wilde, J. C.; Brandes, R. A. C. AdV. Mater. 1993, 5, 466-468. (11) Caminade, A. M.; Majoral, J. P. In Dendrimers, Topics in Current Chemistry; Vo¨gtle, F., Ed.; Springer: Berlin, 1998; pp 79-124. (12) Sournies, F.; Crasnier, F.; Graffeuil, M.; Faucher, J.-P.; Lahana, R.; Labarre, M. C.; Labarre, J.-F. Angew. Chem., Int. Ed. Engl. 1995, 34, 578-581. (13) De Gennes, P.-G.; Hervet, H. J. Phys. Lett. 1983, 44, 351-360. 10.1021/ja021147o CCC: $25.00 © 2003 American Chemical Society

Construction of Giant Dendrimers

ARTICLES

construction, the space AZ available on the dendrimer surface for each terminal group at the dendrimer periphery being given by

AZ ) AD/NZ ∝ r2/NcNbG

(1)

in which AD is the total surface of the periphery, NZ the number of termini, r the radius of the dendrimer sphere, Nc the number of core branches, Nb the multiplication number at each focal point, and G the generation number. De Gennes’ dense-packing (limit) generation Gl follows:

Gl ) a(ln P + b)

(2)

where P is the length of the branch-cell segment, and a and b are parameters that depend on the type of dendrimer (a ) 2.88 and b ) 1.5 for Tomalia’s PAMAM dendrimers1,9). This theory seemingly holds well when the branches are terminated by groups that give hydrogen bonding1 or bear charges,2,3,6 large groups,1-3 or perfluoro groups causing monomer-monomer attraction14 or if the dendrimer has rigid tethers. Besides these cases, it remained essential, however, to clarify whether the dense-packing limit was more generally valid. Other calculations have indeed proposed that the branches turn toward the inside of the dendrimer and that, consequently, the density is larger inside the dendrimer near the core than at the surface.15-17 Viscosimetry,18 solid state NMR,19 13C and 2H NMR relaxation times,20 and fluorescence depolarization21 studies have also suggested that the branches are partly backfolded toward the inside of the dendrimer.6d We report here a dendrimer construction extending beyond the dense-packing limit that is supported inter alia by 1H, 13C, and 29Si NMR spectra recorded after each reaction, indicating that these reactions are clean within the NMR accuracy. Results

The dendritic core and the brick used to construct the dendrimers were synthesized under very mild conditions using high-yield CpFe+-induced perallylation of simple methyl or polymethyl aromatics. Among these known reactions, we have chosen the CpFe+-induced nona-allylation of mesitylene for the synthesis of a nona-allyl core22 and the CpFe+-induced triallylation of p-ethoxytoluene that also proceeds with CpFe+-induced aryl ether cleavage and yields the triallylphenol brick. This simple tripodal compound is now accessible in 60% yield from its easily available precursor [FeCp(η6-p-Cl-C6H4OEt)][PF6].22 We have already reported a tentative divergent dendrimer synthesis using these units.22 It was based on the hydroboration (14) Percec, V.; Johansson, G.; Ungar, G.; Zhou, J. P. J. Am. Chem. Soc. 1996, 18, 9855-9866. (15) Lescanec, R. L.; Muthukumar, M. Macromolecules 1990, 23, 2280-2288. (16) (a) Mansfield, M. L.; Klushin, L. I. Macromolecules 1993, 26, 42624268. (b) Boris, D.; Rubinstein, M. Macromolecules 1996, 29, 7251-7260. (17) Naidoo, K. J.; Hughes, S. J.; Moss, J. R. Macromolecules 1999, 32, 331341. (18) Mourey, T. H.; Turner, S. R.; Rubinstein, M.; Fre´chet, J. M. J.; Hawker, C. J.; Wooley, K. L. Macromolecules 1992, 25, 2401-2406. (19) Wooley, K. L.; Klug, C. A.; Tasaki, K.; Schaefer, J. J. Am. Chem. Soc. 1997, 119, 53-58. (20) Meltzer, A. D.; Tirrell, D. A.; Jones, A. A.; Inglefield, P. T.; Hedstrand, D. M.; Tomalia, D. A. Macromolecules 1992, 25, 4549-4552. (21) De Backer, S.; Prinzie, Y.; Verheijen, W.; Smet, M.; Desmedt, K.; Dehaen, W.; De Schryver, F. C. J. Phys. Chem. A 1998, 102, 5451-5455. (22) Sartor, V.; Djakovitch, L.; Fillaut, J.-L.; Moulines, F.; Neveu, F.; Marvaud, V.; Guittard, J.; Blais, J.-C.; Astruc, D. J. Am. Chem. Soc. 1999, 121, 29292930.

Figure 1. 400 MHz1H and 100 MHz 29Si NMR spectra of the G9-177 047allyl dendrimer, which very clearly show the signals of the surface phenoxy, allyl, and methylene protons and the single 29Si peak.

of the polyallyl core, synthesis of the mesylate, and then nucleophilic substitution of the latter by the phenolate. The yield of the second-generation dendrimer was low because of the steric limit due to the bulk of the large borane groups, which prevented an efficient construction of high generations. In the present work (see Scheme 1), tethers longer than in the previous dendrimers have been designed, which circumvents this problem. The strategy is based first on the hydrosilylation of the polyolefin core catalyzed by the Karsted catalyst in ether using dimethylchloromethylsilane, a reagent already successfully used by Seyferth23 for dendrimer synthesis. The second reaction of the sequence for each generation is the nucleophilic substitution of the terminal chloride by the phenolate group of the triallylphenol catalyzed by sodium iodide in DMF. At each generation, the reaction mixture is passed rapidly through a short column of silica gel after the hydrosilylation in order to remove the catalyst, then the solvent is removed and the residue is submitted without further purification to the second reaction of the sequence. Indeed, the hydrosilylation is virtually quantitative without isomerization according to the accuracy provided by 1H NMR. After the second reaction, the solvent is removed under vacuum, the residue is dissolved in the minimum of methylene chloride, and the dendrimer is precipitated by addition of excess methanol. Then, the gum is dissolved in methylene chloride and submitted to flash chromatography on silica gel, yielding an off-white gum after removing the solvent. Although this procedure lowers the yield, the spectroscopic and analytical data are well optimized in this way. The dendritic construction could be repeated until the ninth generation. Figures 1 and 2 show the 1H and 29Si NMR spectra of the dendrimers G9-177 047 and those of its hydrosilylation product, respectively. The yields were relatively good for the first generations, then progressively decreased (see Tables 1 and 2). At the ninth generation, the yield of purified product became low (8%). The elemental analyses were quite satisfactory until G4-2187, then the carbon content found became deficient for the higher generation numbers (by 1.3-3.4%), implying that impurities were encapsulated inside the dendrimers. The purity was monitored at each generation by 1H and 29Si NMR, and the spectra were always quite clean, showing only the groups present at the periphery of the dendrimers (two last (23) Krsda, S. W.; Seyferth, D. J. Am. Chem. Soc. 1998, 120, 3604-3612. J. AM. CHEM. SOC.

9

VOL. 125, NO. 24, 2003 7251

ARTICLES

Ruiz et al.

Scheme 1. Overall Dendrimer Construction Starting from Ferrocenea

a The nona-allyl core synthesized by CpFe+-induced nona-allylation of mesitylene followed by decomplexation using visible light (right column) can be considered as G0 and thus the 27-allyl derivative, represented below, as G1 (abbreviation: G1-27).

generations, the other generations having too few protons to be observable). From one generation to the next, these spectra remained essentially the same until the ninth generation (Figures 1 and 2). The hydrosilylation reactions were monitored in 1H NMR by the disappearance of the allyl signals at 2.4, 5.0, and 5.5 ppm and the appearance of the sharp singlet corresponding to SiCH2Cl at 2 ppm. The nucleophilic substitutions by the triallylphenate were monitored by the disppearance of this singlet and the increase of the SiCH2O signal at 3.5 ppm and of the phenol signals at 6.9 and 7.2 ppm. The only side reaction that would also lead to a shift from the SiCH2Cl or SiCH2I signal to a SiCH2OR signal different from that with R ) triallylphenol would be that with adventitious water, which would lead to a 7252 J. AM. CHEM. SOC.

9

VOL. 125, NO. 24, 2003

SiCH2OH branch (R ) H). This side reaction can be discarded as follows, however. The mass spectra of the dendrimers G127 and G2-81-allyl do show that such a branch is not formed. If this side reaction would occur for the high generations, the intensity of the signals of the phenol group versus that of the SiCH2OH signal would be weaker than the theoretical one, which is not observed. Finally, the zeroth-generation nonaalcohol dendrimer G0-9-SiCH2OH has been synthesized, and its SiCH2OH methylene signal appears at 3.27 ppm.24 Mixing (24) In accord with our data, a nearly identical 1H NMR chemical shift in CDCl3 (3.22 ppm) for a compound also containing the -SiCH2OH group was recently reported: Tacke, R.; Handmann, V. I. Organometallics 2002, 21, 2619-2625.

Construction of Giant Dendrimers

ARTICLES

Table 1. Characteristics of the Nine Generations of Polyallyl Dendrimers (G1-27 to G9-177 147)

a See Scheme 1 and Figures 3 and 4 for G -G . b Theoretical branch numbers. c See text and Figure 3. d M: monolayer; B: flatenned bilayer. e Molecular 1 3 peak found by MALDI TOF mass spectrometry. f Size-exclusion chromatography, see Figure 5.

Figure 3. MALDI TOF mass spectrum of G2-81-allyl-dendrimer. The major peak is the molecular peak (M ) 11 299; [M + Na]+ at m/z 11 322). The minor peak at 10 422 represents the same dendrimer in which one triallylphenol unit is missing, possibly because of the isomerization of a double bond during the catalyzed hydrosilylation reaction (see text). Figure 2. 400 MHz 1H and 100 MHz 29Si NMR spectra of G9-177 147SiCH2Cl resulting from the catalyzed hydrosilylation of G9-177 147-allyl (see text). See the HRTEM of the iodo analogue in Figure 7. Table 2. Yields (% and mass) Obtained from Generation to Generation and Elemental Analyses (H and C) of the Dendrimers yield from Gn-1 anal. calcd

G1-27-allyl G2-81-allyl G3-243-allyl G4-729-allyl G5-2187-allyl G6-6561-allyl G7-19 683-allyl G8-59 049-allyl G9-177 147-allyl

9.49 9.42 9.40 9.39 9.39 9.39 9.39 9.39

78.05 76.53 76.12 76.00 75.96 75.96 75.94 75.94

anal. found

9.50 9.22 9.60 9.38 9.39 9.66 10.04 9.50

78.25 76.47 76.47 75.42 74.69 73.47 72.62 73.28

%

mass (g)

73 58 54 48 35 28 27 23 8

0.480 0.890 1.530 2.240 2.360 1.980 1.610 1.110 0.270

a In a series of experiments, all the analytical and spectroscopic data were collected (see Experimental Section). In another series, the whole sample (masses in the table) was used for the synthesis of the following generation after checking completion of the reactions by 1H NMR and extensive purification after each reaction as indicated in the text.

it with G2-81-allyl clearly shows both distinct methylene signals due to the two compounds (see the NMR spectra in the Supporting Information). An organic impurity sometimes observed by NMR for the high generations is an alkane residue

(impurity coming from the silica gel). It could be almost completely removed by successive reprecipitation, but a very minor part remained trapped inside the dendrimers. The 29Si spectra were particularly simple. The hydrosilylation gave a compound that showed two 29Si NMR signals at 0.40 and 3.8 ppm, and the polyolefin dendrimers showed only one signal at 0.40 ppm. The dendrimers remained pentane soluble until the ninth generation, and the 1H and 29Si NMR spectra are shown in Figure 1. Although the impurities attached to the dendrimers cannot be seen within the NMR accuracy (